Mehdi Khribch (المهدي خريبش)
Postdoctoral Researcher in Statistics & Machine Learning
About Me
I am a postdoctoral researcher funded by the ANR project BACKUP. I am based between the Laboratoire de Mathématiques d'Orsay (Université Paris-Saclay) and the LPSM (Sorbonne Université and Université Paris Cité), working with Ismaël Castillo, Gilles Blanchard and Guillermo Durand. Until recently, I was a PhD candidate at ESSEC Business School in Paris, within the Department of Information Systems, Data Analytics and Operations, advised by Pierre E. Jacob and Pierre Alquier. I hold an Engineering Degree from Mines Paris—PSL University, the Master MVA (Mathématiques, Vision, Apprentissage) from ENS Paris-Saclay, and an MSc in Statistical Science from the Department of Statistics, University of Oxford. Outside of research, I am passionate about literature, cinema, and art. See my blog to learn more.
Research Interests
My research lies at the intersection of Monte Carlo methods, statistical learning theory, and high-dimensional statistics, in particular Bayesian nonparametrics, variational inference and Bayesian deep learning. Much of my work so far has been on importance sampling and MCMC, especially global proposal Monte Carlo methods and the analysis of their subgeometric convergence regimes. I am also interested in generalization guarantees, PAC-Bayesian theory, and robustness to model misspecification. These days, I am increasingly drawn to the question of how priors can adapt to the unknown structure of the data, from nonparametric models to Bayesian neural networks. The common thread is the development of scalable, theoretically principled algorithms for statistical learning.
Feel free to get in touch.
